{"id":"W4408007200","doi":"10.1101/2025.02.27.25323003","title":"Accuracy of preferred language data in a multi-hospital electronic health record in Toronto, Canada","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Interpreting and Communication in Healthcare","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University Health Network; Sunnybrook Health Science Centre; Sinai Health System; The Scarborough Hospital","funders":"","keywords":"Electronic health record; Geography; Computer science; Data science; Medicine; Health care; Economics; Economic growth","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.004556959,0.0002803264,0.0003728594,0.00224621,0.001282004,0.002143365,0.00129965,0.0002827257,0.001212926],"category_scores_gemma":[0.03459017,0.0002994401,0.0003939906,0.00483487,0.0006532808,0.0005908656,0.001206453,0.0004224497,0.0002230201],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.03116058,"about_ca_system_score_gemma":0.03304969,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9777434,"about_ca_topic_score_gemma":0.9842296,"domain_scores_codex":[0.9899495,0.002010967,0.001609536,0.0009017912,0.004667897,0.0008603401],"domain_scores_gemma":[0.9396167,0.008807608,0.01049334,0.002058826,0.03677322,0.002250248],"domain_codex":null,"domain_gemma":"reporting","domain_candidate":"reporting","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002328556,0.00002378936,0.9754205,0.000230159,0.00007541102,0.0002299131,0.002318486,0.0007135147,0.0004438645,0.0001988023,0.002214144,0.01789844],"study_design_scores_gemma":[0.00001436325,0.00004532766,0.9895752,0.0002536331,0.00004595075,0.0001359724,0.002747886,0.003931338,0.0007051408,0.00005846756,0.002454496,0.00003220901],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9813139,0.001662051,0.001288956,0.00165159,0.00004050526,0.0001355532,0.009249932,0.0000837079,0.004573894],"genre_scores_gemma":[0.9955477,0.0004364692,0.001239424,0.0001684384,0.000009084601,0.00002386749,0.001953185,0.000009863912,0.0006120704],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.995443,"threshold_uncertainty_score":0.2260869,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08412317312320126,"score_gpt":0.4709129388863321,"score_spread":0.3867897657631308,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}